Impact of Workspace Design on Radiation Therapist Technicians' Physical Stressors, Mental Workload, Situation Awareness, and Performance.

PURPOSE: Our purpose was to assess the effect of workspace configuration on radiation therapists' (RTs) physical stressors, mental workload (WL), situational awareness (SA), and performance during routine treatment delivery tasks in a simulated environment. METHODS AND MATERIALS: Fourteen RTs were randomized to 2 workspace configurations while performing 4 simulated scenarios: current (not ergonomically optimized; n = 7) and enhanced (ergonomically optimized, n = 7). Physical stressors were objectively assessed using a rapid upper limb assessment tool. Mental WL was measured at the end of each simulated scenario subjectively using the NASA Task-Load Index and objectively throughout the scenario using eye-tracking metrics (pupil diameter and blink rate). SA was measured at the end of each simulated scenario subjectively using the situation awareness and review technique. Performance was measured objectively via assessment of time-out compliance, error detection, and procedural compliance. Analysis of variance was used to test the effect of workspace configuration on physical stressors, mental WL, SA, and performance. RESULTS: The enhanced configuration significantly reduced physical stressors (rapid upper limb assessment; P < .01) and resulted in a higher rate of time-out compliance (P = .01) compared with current workspace configuration. No significant effect on other metrics was measured. CONCLUSIONS: Our results suggest that an ergonomically designed workspace may minimize physical stressors and improve the performance of RTs.

Predicting Radiation Therapy Process Reliability Using Voluntary Incident Learning System Data.

PURPOSE: This study aimed to present an innovative approach to quantify, visualize, and predict radiation therapy (RT) process reliability using data captured from a voluntary incident learning system, with an overall aim to improve patient safety outcomes. METHODS AND MATERIALS: We analyzed 111 reported deviations that were tripped and caught within 159 mapped RT process steps included within 7 major stages of RT delivery, 94 of which were any type of quality assurance (QA) controls. This allowed for us to compute the trip rate and fail-to-catch-rate (FCR) per each QA control with the available data. Next, we used a logistic regression model to identify significant variables predictive of FCRs, predicted FCRs for each QA control without available data, and thus, attempted to quantify RT process reliability expressed as percentage of patients with uncaught deviations after treatment planning, before their first treatment, and during treatment delivery. RESULTS: Using the predicted FCRs, we computed the upper 95% likelihood that a deviation remains uncaught in a patient's course of treatment at the following RT process stages: immediately after treatment planning at 10.26%; before the first treatment at 0.0052%; and throughout treatment delivery at 0.0276%. CONCLUSIONS: The results suggest that RT process reliability can be predicted and visualized using data from incident learning systems. If implemented and used as a safety metric, this could help RT clinics to proactively maintain their preoccupation with patient safety. RT process reliability may also help guide future work on standardization and continuous improvement of the design of RT QA programs.

Promoting safety mindfulness: Recommendations for the design and use of simulation-based training in radiation therapy.

There is a need to better prepare radiation therapy (RT) providers to safely operate within the health information technology (IT) sociotechnical system. Simulation-based training has been preemptively used to yield meaningful improvements during providers' interactions with health IT, including RT settings. Therefore, on the basis of the available literature and our experience, we propose principles for the effective design and use of simulated scenarios and describe a conceptual framework for a debriefing approach to foster successful training that is focused on safety mindfulness during RT professionals' interactions with health IT.

Developing and assessing electronic checklists for safety mindfulness, workload, and performance.

PURPOSE: The aim of this study is to propose a set of innovative principles for the effective design of electronic checklists to enhance safety mindfulness (a specific safety mindful mindset that offers the opportunity to operate more preemptively during routine quality assurance tasks) and discuss some of our preliminary results from testing our proposed electronic checklist with dosimetrists and physicists. METHODS AND MATERIALS: A multidisciplinary team designed, developed, and evaluated the utility of the electronic checklist (vs paper-based checklist) to promote safety mindfulness. Subjective workload was measured at the end of each assessment/scenario. Performance was quantified on the basis of discovery of purposefully embedded errors, time to complete the scenario, and additional concerns that were documented by the participants. RESULTS: Use of the electronic checklist was associated with decreases in time to scenario completion (P < .01) and increases in documentation of additional patient safety and plan quality concerns (P = .04) but had no significant impact on the recognition of purposefully embedded errors or perceptions of workload. CONCLUSIONS: Our proposed principles for the design of electronic checklists may improve the efficiency of quality assurance procedures while enhancing users' safety mindfulness. Future research is needed to better understand the utility of our proposed design principles on patient safety from a long-term use perspective.

Identifying Factors and Root Causes Associated With Near-Miss or Safety Incidents in Patients Treated With Radiotherapy: A Case-Control Analysis.

PURPOSE: To identify factors associated with a near-miss or safety incident (NMSI) in patients undergoing radiotherapy and identify common root causes of NMSIs and their relationship with incident severity. METHODS: We retrospectively studied NMSIs filed between October 2014 and April 2016. We extracted patient-, treatment-, and disease-specific data from patients with an NMSI (n = 200; incident group) and a similar group of control patients (n = 200) matched in time, without an NMSI. A root cause and incident severity were determined for each NMSI. Univariable and multivariable analyses were performed to determine which specific factors were contributing to NMSIs. Multivariable logistic regression was used to determine root causes of NMSIs and their relationship with incident severity. RESULTS: NMSIs were associated with the following factors: head and neck sites (odds ratio [OR], 5.2; P = .01), image-guided intensity-modulated radiotherapy (OR, 3; P = .009), daily imaging (OR, 7; P < .001), and tumors staged as T2 (OR, 3.3; P = .004). Documentation and scheduling errors were the most common root causes (29%). Communication errors were more likely to affect patients ( P < .001), and technical treatment delivery errors were most associated with a higher severity score ( P = .005). CONCLUSION: Several treatment- and disease-specific factors were found to be associated with an NMSI. Overall, our results suggest that complexity (eg, head and neck, image-guided intensity-modulated radiotherapy, and daily imaging) might be a contributing factor for an NMSI. This promotes an idea of developing a more dedicated and robust quality assurance system for complex cases and highlights the importance of a strong reporting system to support a safety culture.

Improving radiation oncology providers' workload and performance: Can simulation-based training help?

PURPOSE: To help with ongoing safety challenges in radiation therapy (RT), the objective of this research was to develop and assess the impact of a simulation-based training intervention on radiation oncology providers' workload and performance during treatment planning and quality assurance (QA) tasks. METHODS AND MATERIALS: Eighteen radiation oncology professionals completed routine treatment planning and QA tasks on 2 clinical scenarios in a simulation laboratory as part of a prospective institutional review board-approved study. Workload was measured at the end of each assessment/scenario using the NASA Task-Load Index. Performance was quantified based on procedural compliance (adherence to preset/standard QA tasks), time-to-scenario completion, and clinically relevant performance. Participants were then randomized to receive (vs not receive) simulation-based training intervention (eg, standardized feedback on workload and performance) and underwent repeat measurements of workload and performance. Pre- and postintervention changes in workload and performance from participants who received (vs did not receive) were compared using 2-way analysis of variance. RESULTS: Simulation-based training was associated with significant improvements in procedural compliance (P = .01) and increases in time-to-scenario completion (P < .01) but had no significant impact on subjective workload or clinically relevant performance. CONCLUSION: Simulation-based training may be a tool to improve procedural compliance of RT professionals and to acquire new skills and knowledge to proactively maintain RT professionals' preoccupation with patient safety.

Toward a better understanding of task demands, workload, and performance during physician-computer interactions.

OBJECTIVE: To assess the relationship between (1) task demands and workload, (2) task demands and performance, and (3) workload and performance, all during physician-computer interactions in a simulated environment. METHODS: Two experiments were performed in 2 different electronic medical record (EMR) environments: WebCIS (n = 12) and Epic (n = 17). Each participant was instructed to complete a set of prespecified tasks on 3 routine clinical EMR-based scenarios: urinary tract infection (UTI), pneumonia (PN), and heart failure (HF). Task demands were quantified using behavioral responses (click and time analysis). At the end of each scenario, subjective workload was measured using the NASA-Task-Load Index (NASA-TLX). Physiological workload was measured using pupillary dilation and electroencephalography (EEG) data collected throughout the scenarios. Performance was quantified based on the maximum severity of omission errors. RESULTS: Data analysis indicated that the PN and HF scenarios were significantly more demanding than the UTI scenario for participants using WebCIS (P < .01), and that the PN scenario was significantly more demanding than the UTI and HF scenarios for participants using Epic (P < .01). In both experiments, the regression analysis indicated a significant relationship only between task demands and performance (P < .01). DISCUSSION: Results suggest that task demands as experienced by participants are related to participants' performance. Future work may support the notion that task demands could be used as a quality metric that is likely representative of performance, and perhaps patient outcomes. CONCLUSION: The present study is a reasonable next step in a systematic assessment of how task demands and workload are related to performance in EMR-evolving environments.

Use of mobile device technology to continuously collect patient-reported symptoms during radiation therapy for head and neck cancer: A prospective feasibility study.

PURPOSE: Accurate assessment of toxicity allows for timely delivery of supportive measures during radiation therapy for head and neck cancer. The current paradigm requires weekly evaluation of patients by a provider. The purpose of this study is to evaluate the feasibility of monitoring patient reported symptoms via mobile devices. METHODS AND MATERIALS: We developed a mobile application for patients to report symptoms in 5 domains using validated questions. Patients were asked to report symptoms using a mobile device once daily during treatment or more often as needed. Clinicians reviewed patient-reported symptoms during weekly symptom management visits and patients completed surveys regarding perceptions of the utility of the mobile application. The primary outcome measure was patient compliance with mobile device reporting. Compliance is defined as number of days with a symptom report divided by number of days on study. RESULTS: There were 921 symptom reports collected from 22 patients during treatment. Median reporting compliance was 71% (interquartile range, 45%-80%). Median number of reports submitted per patient was 34 (interquartile range, 21-53). Median number of reports submitted by patients per week was similar throughout radiation therapy and there was significant reporting during nonclinic hours. Patients reported high satisfaction with the use of mobile devices to report symptoms. CONCLUSIONS: A substantial percentage of patients used mobile devices to continuously report symptoms throughout a course of radiation therapy for head and neck cancer. Future studies should evaluate the impact of mobile device symptom reporting on improving patient outcomes.

The association between event learning and continuous quality improvement programs and culture of patient safety.

PURPOSE: To present our approach and results from our quality and safety program and to report their possible impact on our culture of patient safety. METHODS AND MATERIALS: We created an event learning system (termed a "good catch" program) and encouraged staff to report any quality or safety concerns in real time. Events were analyzed to assess the utility of safety barriers. A formal continuous quality improvement program was created to address these reported events and make improvements. Data on perceptions of the culture of patient safety were collected using the Agency for Health Care Research and Quality survey administered before, during, and after the initiatives. RESULTS: Of 560 good catches reported, 367 could be ascribed to a specific step on our process map. The calculated utility of safety barriers was highest for those embedded into the pretreatment quality assurance checks performed by physicists and dosimetrists (utility score 0.53; 93 of 174) and routine checks done by therapists on the initial day of therapy. Therapists and physicists reported the highest number of good catches (24% each). Sixty-four percent of events were caused by performance issues (eg, not following standardized processes, including suboptimal communications). Of 31 initiated formal improvement events, 26 were successfully implemented and sustained, 4 were discontinued, and 1 was not implemented. Most of the continuous quality improvement program was conducted by nurses (14) and therapists (7). Percentages of positive responses in the patient safety culture survey appear to have increased on all dimensions (p < .05). CONCLUSIONS: Results suggest that event learning and continuous quality improvement programs can be successfully implemented and that there are contemporaneous improvements in the culture of safety.

Direct aperture optimization using an inverse form of back-projection.

Direct aperture optimization (DAO) has been used to produce high dosimetric quality intensity-modulated radiotherapy (IMRT) treatment plans with fast treatment delivery by directly modeling the multileaf collimator segment shapes and weights. To improve plan quality and reduce treatment time for our in-house treatment planning system, we implemented a new DAO approach without using a global objective function (GFO). An index concept is introduced as an inverse form of back-projection used in the CT multiplicative algebraic reconstruction technique (MART). The index, introduced for IMRT optimization in this work, is analogous to the multiplicand in MART. The index is defined as the ratio of the optima over the current. It is assigned to each voxel and beamlet to optimize the fluence map. The indices for beamlets and segments are used to optimize multileaf collimator (MLC) segment shapes and segment weights, respectively. Preliminary data show that without sacrificing dosimetric quality, the implementation of the DAO reduced average IMRT treatment time from 13 min to 8 min for the prostate, and from 15 min to 9 min for the head and neck using our in-house treatment planning system PlanUNC. The DAO approach has also shown promise in optimizing rotational IMRT with burst mode in a head and neck test case.

Clinical experience with 3-dimensional surface matching-based deep inspiration breath hold for left-sided breast cancer radiation therapy.

PURPOSE: Three-dimensional (3D) surface matching is a novel method to administer deep inspiration breath-hold (DIBH) radiation therapy for left-sided breast cancer to reduce cardiac exposure. We analyzed port (x-ray) films to assess patient setup accuracy and treatment times to assess the practical workflow of this system. METHODS AND MATERIALS: The data from 50 left-sided breast cancer patients treated with DIBH were studied. AlignRT (London, UK) was used. The distance between the field edge and the anterior pericardial shadow as seen on the routine port films (dPORT), and the corresponding distance seen on the digitally reconstructed radiographs (DRR) from the planning (dDRR) were compared as a quantitative measure of setup accuracy. Variations of dPORT - dDRR over the treatment course were assessed. In a subset of 21 patients treated with tangential beams alone, the daily treatment durations were analyzed to assess the practical workflow of this system. RESULTS: Considering all 50 patients, the mean absolute systematic uncertainty between dPORT and dDRR was 0.20 cm (range, 0 to 1.22 cm), the mean systematic uncertainty was -0.07 cm (range, -1.22 to 0.67 cm), and their mean random uncertainty was 0.19 cm (range, 0 to 0.84 cm). There was no significant change in dPORT - dDRR during the course of treatment. The mean patient treatment duration for the 21 patients studied was 11 minutes 48 seconds. On intrapatient assessments, 15/21 had nonsignificant trends toward reduced treatment durations during their course of therapy. On interpatient comparisons, the mean treatment times declined as we gained more experience with this technique. CONCLUSIONS: The DIBH patient setup appears to provide a fairly reproducible degree of cardiac sparing with random uncertainties of ≈ 0.2 cm. The treatment durations are clinically acceptable and appear not to change significantly over time on an intrapatient basis, and to improve over time on an interpatient basis.

Comparison of User-Directed and Automatic Mapping of the Planned Isocenter to Treatment Space for Prostate IGRT.

Image-guided radiotherapy (IGRT), adaptive radiotherapy (ART), and online reoptimization rely on accurate mapping of the radiation beam isocenter(s) from planning to treatment space. This mapping involves rigid and/or nonrigid registration of planning (pCT) and intratreatment (tCT) CT images. The purpose of this study was to retrospectively compare a fully automatic approach, including a non-rigid step, against a user-directed rigid method implemented in a clinical IGRT protocol for prostate cancer. Isocenters resulting from automatic and clinical mappings were compared to reference isocenters carefully determined in each tCT. Comparison was based on displacements from the reference isocenters and prostate dose-volume histograms (DVHs). Ten patients with a total of 243 tCTs were investigated. Fully automatic registration was found to be as accurate as the clinical protocol but more precise for all patients. The average of the unsigned x, y, and z offsets and the standard deviations ( σ ) of the signed offsets computed over all images were (avg. ± σ (mm)): 1.1 ± 1.4, 1.8 ± 2.3, 2.5 ± 3.5 for the clinical protocol and 0.6 ± 0.8, 1.1 ± 1.5 and 1.1 ± 1.4 for the automatic method. No failures or outliers from automatic mapping were observed, while 8 outliers occurred for the clinical protocol.

The impact of local and regional disease extent on overall survival in patients with advanced stage IIIB/IV non-small cell lung carcinoma.

PURPOSE: Patients with advanced stage IIIB or stage IV non-small cell lung carcinoma are typically treated with initial platinum-based chemotherapy. A variety of factors (eg, performance status, gender, age, histology, weight loss, and smoking history) are generally accepted as predictors of overall survival. Because uncontrolled pulmonary disease constitutes a major cause of death in these patients, we hypothesized that clinical and radiographic factors related to intrathoracic disease at diagnosis may be prognostically significant in addition to conventional factors. The results have implications regarding the selection of patients for whom palliative thoracic radiation therapy may be of most benefit. METHODS AND MATERIALS: We conducted a pooled analysis of 189 patients enrolled at a single institution into 9 prospective phase II and III clinical trials involving first-line, platinum-based chemotherapy. Baseline clinical and radiographic characteristics before trial enrollment were analyzed as possible predictors for subsequent overall survival. To assess the relationship between anatomic location and volume of disease within the thorax and its effect on survival, the pre-enrollment computed tomography images were also analyzed by contouring central and peripheral intrapulmonary disease. RESULTS: On univariate survival analysis, multiple pulmonary-related factors were significantly associated with worse overall survival, including pulmonary symptoms at presentation (P=.0046), total volume of intrathoracic disease (P=.0006), and evidence of obstruction of major bronchi or vessels on prechemotherapy computed tomography (P<.0001). When partitioned into central and peripheral volumes, central (P<.0001) but not peripheral (P=.74) disease was associated with worse survival. On multivariate analysis with known factors, pulmonary symptoms (hazard ratio, 1.46; P=.042), central disease volume (hazard ratio, 1.47; P=.042), and bronchial/vascular compression (hazard ratio, 1.54; P=.022) remained significant. CONCLUSIONS: Patients with bulky central disease, bronchial/vascular compression, and/or pulmonary symptoms exhibited worse overall survival after first-line, platinum-based chemotherapy. A subset of these patients may be studied to determine whether early, planned palliative thoracic radiation could also be of benefit.

Training models of anatomic shape variability.

Learning probability distributions of the shape of anatomic structures requires fitting shape representations to human expert segmentations from training sets of medical images. The quality of statistical segmentation and registration methods is directly related to the quality of this initial shape fitting, yet the subject is largely overlooked or described in an ad hoc way. This article presents a set of general principles to guide such training. Our novel method is to jointly estimate both the best geometric model for any given image and the shape distribution for the entire population of training images by iteratively relaxing purely geometric constraints in favor of the converging shape probabilities as the fitted objects converge to their target segmentations. The geometric constraints are carefully crafted both to obtain legal, nonself-interpenetrating shapes and to impose the model-to-model correspondences required for useful statistical analysis. The paper closes with example applications of the method to synthetic and real patient CT image sets, including same patient male pelvis and head and neck images, and cross patient kidney and brain images. Finally, we outline how this shape training serves as the basis for our approach to IGRT/ART.

A method and software for segmentation of anatomic object ensembles by deformable m-reps.

Deformable shape models (DSMs) comprise a general approach that shows great promise for automatic image segmentation. Published studies by others and our own research results strongly suggest that segmentation of a normal or near-normal object from 3D medical images will be most successful when the DSM approach uses (1) knowledge of the geometry of not only the target anatomic object but also the ensemble of objects providing context for the target object and (2) knowledge of the image intensities to be expected relative to the geometry of the target and contextual objects. The segmentation will be most efficient when the deformation operates at multiple object-related scales and uses deformations that include not just local translations but the biologically important transformations of bending and twisting, i.e., local rotation, and local magnification. In computer vision an important class of DSM methods uses explicit geometric models in a Bayesian statistical framework to provide a priori information used in posterior optimization to match the DSM against a target image. In this approach a DSM of the object to be segmented is placed in the target image data and undergoes a series of rigid and nonrigid transformations that deform the model to closely match the target object. The deformation process is driven by optimizing an objective function that has terms for the geometric typicality and model-to-image match for each instance of the deformed model. The success of this approach depends strongly on the object representation, i.e., the structural details and parameter set for the DSM, which in turn determines the analytic form of the objective function. This paper describes a form of DSM called m-reps that has or allows these properties, and a method of segmentation consisting of large to small scale posterior optimization of m-reps. Segmentation by deformable m-reps, together with the appropriate data representations, visualizations, and user interface, has been implemented in software that accomplishes 3D segmentations in a few minutes. Software for building and training models has also been developed. The methods underlying this software and its abilities are the subject of this paper.

Comparison of human and automatic segmentations of kidneys from CT images.

PURPOSE: A controlled observer study was conducted to compare a method for automatic image segmentation with conventional user-guided segmentation of right and left kidneys from planning computerized tomographic (CT) images. METHODS AND MATERIALS: Deformable shape models called m-reps were used to automatically segment right and left kidneys from 12 target CT images, and the results were compared with careful manual segmentations performed by two human experts. M-rep models were trained based on manual segmentations from a collection of images that did not include the targets. Segmentation using m-reps began with interactive initialization to position the kidney model over the target kidney in the image data. Fully automatic segmentation proceeded through two stages at successively smaller spatial scales. At the first stage, a global similarity transformation of the kidney model was computed to position the model closer to the target kidney. The similarity transformation was followed by large-scale deformations based on principal geodesic analysis (PGA). During the second stage, the medial atoms comprising the m-rep model were deformed one by one. This procedure was iterated until no changes were observed. The transformations and deformations at both stages were driven by optimizing an objective function with two terms. One term penalized the currently deformed m-rep by an amount proportional to its deviation from the mean m-rep derived from PGA of the training segmentations. The second term computed a model-to-image match term based on the goodness of match of the trained intensity template for the currently deformed m-rep with the corresponding intensity data in the target image. Human and m-rep segmentations were compared using quantitative metrics provided in a toolset called Valmet. Metrics reported in this article include (1) percent volume overlap; (2) mean surface distance between two segmentations; and (3) maximum surface separation (Hausdorff distance). RESULTS: Averaged over all kidneys the mean surface separation was 0.12 cm, the mean Hausdorff distance was 0.99 cm, and the mean volume overlap for human segmentations was 88.8%. Between human and m-rep segmentations the mean surface separation was 0.18-0.19 cm, the mean Hausdorff distance was 1.14-1.25 cm, and the mean volume overlap was 82-83%. CONCLUSIONS: Overall in this study, the best m-rep kidney segmentations were at least as good as careful manual slice-by-slice segmentations performed by two experienced humans, and the worst performance was no worse than typical segmentations from our clinical setting. The mean surface separations for human-m-rep segmentations were slightly larger than for human-human segmentations but still in the subvoxel range, and volume overlap and maximum surface separation were slightly better for human-human comparisons. These results were expected because of experimental factors that favored comparison of the human-human segmentations. In particular, m-rep agreement with humans appears to have been limited largely by fundamental differences between manual slice-by-slice and true three-dimensional segmentation, imaging artifacts, image voxel dimensions, and the use of an m-rep model that produced a smooth surface across the renal pelvis.

Beam orientation selection for intensity-modulated radiation therapy based on target equivalent uniform dose maximization.

PURPOSE: To develop an automated beam-orientation selection procedure for intensity-modulated radiotherapy (IMRT), and to determine if a small number of beams picked by this automated procedure can yield results comparable to a large number of manually placed orientations. METHODS AND MATERIALS: The automated beam selection procedure maximizes an unconstrained objective function composed of target equivalent uniform dose (EUD) and critical structure dose-volume histogram (DVH) constraints. Beam orientations are selected from a large feasible set of directions through a series of alternating fluence optimization and orientation alteration steps, until convergence to a stable orientation set. The fluence optimization step adjusts fluences to maximize the objective function. The orientation alteration step substitutes beams in the orientation set currently under consideration with beams of the parent set in the immediate angular vicinity; the altered orientation set is deemed current if it produces a higher objective function value in the fluence optimization step. RESULTS AND CONCLUSIONS: It is demonstrated, for prostate IMRT planning, that a modest number of appropriately selected beam orientations (3 or 5) can provide dose distributions as satisfactory as those produced by a large number of unselected equispaced orientations. Such selected beam orientations can reduce overall treatment time, thus making IMRT more clinically practical.

Deformable M-Reps for 3D Medical Image Segmentation.

M-reps (formerly called DSLs) are a multiscale medial means for modeling and rendering 3D solid geometry. They are particularly well suited to model anatomic objects and in particular to capture prior geometric information effectively in deformable models segmentation approaches. The representation is based on figural models, which define objects at coarse scale by a hierarchy of figures - each figure generally a slab representing a solid region and its boundary simultaneously. This paper focuses on the use of single figure models to segment objects of relatively simple structure. A single figure is a sheet of medial atoms, which is interpolated from the model formed by a net, i.e., a mesh or chain, of medial atoms (hence the name m-reps), each atom modeling a solid region via not only a position and a width but also a local figural frame giving figural directions and an object angle between opposing, corresponding positions on the boundary implied by the m-rep. The special capability of an m-rep is to provide spatial and orientational correspondence between an object in two different states of deformation. This ability is central to effective measurement of both geometric typicality and geometry to image match, the two terms of the objective function optimized in segmentation by deformable models. The other ability of m-reps central to effective segmentation is their ability to support segmentation at multiple levels of scale, with successively finer precision. Objects modeled by single figures are segmented first by a similarity transform augmented by object elongation, then by adjustment of each medial atom, and finally by displacing a dense sampling of the m-rep implied boundary. While these models and approaches also exist in 2D, we focus on 3D objects. The segmentation of the kidney from CT and the hippocampus from MRI serve as the major examples in this paper. The accuracy of segmentation as compared to manual, slice-by-slice segmentation is reported.

Thresholds for human detection of patient setup errors in digitally reconstructed portal images of prostate fields.

PURPOSE: Computer-assisted methods to analyze electronic portal images for the presence of treatment setup errors should be studied in controlled experiments before use in the clinical setting. Validation experiments using images that contain known errors usually report the smallest errors that can be detected by the image analysis algorithm. This paper offers human error-detection thresholds as one benchmark for evaluating the smallest errors detected by algorithms. Unfortunately, reliable data are lacking describing human performance. The most rigorous benchmarks for human performance are obtained under conditions that favor error detection. To establish such benchmarks, controlled observer studies were carried out to determine the thresholds of detectability for in-plane and out-of-plane translation and rotation setup errors introduced into digitally reconstructed portal radiographs (DRPRs) of prostate fields. METHODS AND MATERIALS: Seventeen observers comprising radiation oncologists, radiation oncology residents, physicists, and therapy students participated in a two-alternative forced choice experiment involving 378 DRPRs computed using the National Library of Medicine Visible Human data sets. An observer viewed three images at a time displayed on adjacent computer monitors. Each image triplet included a reference digitally reconstructed radiograph displayed on the central monitor and two DRPRs displayed on the flanking monitors. One DRPR was error free. The other DRPR contained a known in-plane or out-of-plane error in the placement of the treatment field over a target region in the pelvis. The range for each type of error was determined from pilot observer studies based on a Probit model for error detection. The smallest errors approached the limit of human visual capability. The observer was told what kind of error was introduced, and was asked to choose the DRPR that contained the error. Observer decisions were recorded and analyzed using repeated-measures analysis of variance. RESULTS: The thresholds of detectability averaged over all observers were approximately 2.5 mm for in-plane translations, 1.6 degrees for in-plane rotations, 1 degrees for out-of-plane rotations, and 8% change in magnification for out-of-plane translations along the central axis. When one inexperienced observer is excluded, the average threshold for change in magnification is 5%. Experienced observers tended to perform better, but differences between groups were not statistically significant. Thresholds were computed as averages over all observers. Because of the broad range of observer capabilities, some detection tasks were too difficult for some observers, leading to missing threshold values in our data analysis. The missing values were excluded from computation of the average thresholds reported above. The effect of the missing values is to bias the average values toward the best human performance. CONCLUSIONS: Under favorable conditions, humans can detect small errors in setup geometry. The thresholds for error detection reported in this study are believed to represent rigorous but reasonable benchmarks that can be incorporated into studies evaluating algorithms for computer-assisted detection of setup errors in electronic portal images.